single molecule imaging of oxygenation of cobalt ...single molecule imaging of oxygenation of cobalt...
TRANSCRIPT
1
"This document is the Accepted Manuscript version of a Published Work that appeared in final form in J. Amer. Chem. Soc., copyright © American Chemical Society after peer review and technical editing by the publisher. To access the final edited and published work see J. Am.
Chem. Soc. 2012, 134, 14897−14904
Single Molecule Imaging of Oxygenation of Cobalt Octaethylporphyrin
at the Solution/Solid Interface: Thermodynamics from Microscopy.
Benjamin A. Friesen, Ashish Bhattarai, Ursula Mazur*, and K. W. Hipps*
Department of Chemistry
Washington State University
Pullman, WA 99164-4630
ABSTRACT
For the first time, the pressure and temperature dependence of a chemical reaction at the solid
solution interface is studied by STM and thermodynamic data is derived. In particular, the STM
is used to study the reversible binding of O2 with cobalt(II) octaethylporphyrin (CoOEP)
supported on highly oriented pyrolytic graphite (HOPG) at the phenyloctane/CoOEP/HOPG
interface. The adsorption is shown to follow the Langmuir isotherm with P1/2298K
= 3,200 Torr.
Over the temperature range of 10ºC to 40ºC it was found that HP = -68±10 kJ/mol and SP =
-297±30 J/(mol K). The enthalpy and entropy changes are slightly larger than expected based on
solution phase reactions, and possible origins of these differences are discussed. The big surprise
here is the presence of any O2 binding at room temperature, since CoOEP is not expected to bind
O2 in fluid solution. The stability of the bound oxygen is attributed to charge donation from the
graphite substrate to the cobalt, thereby stabilizing the polarized Co-O2 bonding. We report the
surface unit cell for CoOEP on HOPG in phenyloctane at 25C to be A = (1.46±0.1)n nm, B =
(1.36±0.1)m nm, and = 54±3º, where n and m are unknown non-zero non-negative integers.
2
INTRODUCTION
Temperature dependent studies of surface structures can provide a great wealth of
information. They can yield diffusion rates, reaction rates, activation energies, thermodynamic
quantities such as entropy and enthalpy of adsorption and/or surface reaction. Because different
surface species reach equilibrium at different temperatures, and because some surface reactions
are kinetically controlled, a study of a given solution-surface pair as a function of temperature
can lead to the discovery of new materials and phases. In some cases, the ability to even observe
the surface structure is confined to a relatively narrow temperature window, while in others,
different structures are observed over a wide temperature range. The ability to observe surface
structures as a function of temperature at the solution-solid interface has particular relevance to
modern technology, self-assembly, catalysis, friction, film growth, organic electronics, and many
other areas.
The study of molecular and even atomic processes at surfaces has been dramatically
advanced by the application of the scanning tunneling microscope (STM). There are now about
27,000 papers having “scanning tunneling microscopy” in the title or abstract. If one narrows a
SciFinder search to include the "solution interface", the number of citations drops to about 500,
but still a massive body of literature. There are many tens of papers where samples are
separately heated, returned to ambient, and then measured. If one then searches for those papers
where molecular level studies are actually performed at the solid-solution interface at a
temperature significantly different from room temperature, the total publications drop to a very
few, 1-8 with many of these at a single fixed temperature and at least one of the very recent papers
inspired by our recent work. 9 A few more papers focus on hot stage designs appropriate for this
application (but often focused on AFM applications).10-13
3
In this work we will use variable temperature STM imaging at the solution solid interface
to determine the thermodynamics of the reversible binding of dioxygen to CoOEP adsorbed on
highly oriented pyrolytic graphite (HOPG). Both the variable temperature and single molecule
resolution capabilities of the scanning tunneling microscope (STM) are critical for this study. To
our knowledge this the first single molecule level study of the temperature dependence of a
chemical reaction at the solution-solid interface of any system, not just the dioxygen-CoOEP
system. The STM offers unique features including sub-molecular resolution, sensitivity to
electronic structure, ability to function at the solution-solid interface, and variable temperature
capability that make it an ideal tool for the study of temperature dependent surface reactions.
Metalloporphyrins and metallophthalocyanines are a highly versatile family of molecules
with widely varying properties and potential applications. Porphyrins and phthalocyanines may
serve as components in solar cells,14 fuel cells,15 non-linear optical devices,16 sensors,17
catalysts,18 serve as sensitizers in photodynamic tumor therapy,19 and as components in
nanostructured materials.20,21 The binding of small molecules to metalloporphyrins and
metallophthalocyanines is a topic of great interest for a number of reasons. Metalloporphyrins
are analogs of hemes, and consequently knowledge of the binding of dioxygen to
metalloporphyrins is critical for our understanding of life.22 Several gas sensing systems
employing metallated porphyrins and phthalocyanines have been reported in the literature.
Cobalt phthalocyanine molecules deposited on gold electrodes via organic molecular beam
epitaxy show promise as sensors for a variety of species including the Sarin analog dimethyl
methylphosphonate.23,24 The cobalt “picket fence” porphyrin, meso-α,α,α,α-tetrakis(o-
pivalamidophenyl) porphyrinatocobalt(II) imbedded in a polymer matrix functions as an oxygen
sensor.25 Cobalt porphyrins also show promise as electrocatalysts for oxygen reduction.26,27,28,29
4
While the binding of dioxygen to cobalt porphyrins is well known much of the data reported
was gathered at low temperatures (below -20°C) either in solution30-37 or in frozen solutions.38,39
There are relatively few cobalt porphyrins capable of reversibly binding dioxygen at ambient
temperatures,40 most notably cobalt substituted myoglobins41,42 and the aforementioned cobalt
“picket fence” porphyrin meso-α,α,α,α-tetrakis(o-pivalamidophenyl) porphyinatocobalt(II). This
porphyrin binds dioxygen at room temperature in solution40-47 and in the solid state,45 but only
when a basic axial ligand is present. Kinetic data on the adsorption/desorption of dioxygen to
this molecule have been reported at 40°C, wherein a 2-stage adsorption model was proposed.44
The enthalpy and entropy changes associated with dioxygen binding to cobalt substituted
porphyrins are of interest because cobalt porphyrins serve as model systems for the study of
oxygen binding to hemes. Much of the compiled thermodynamic data on cobalt substituted
naturally occurring and model porphyrins has been gathered in solution with enthalpy and
entropy values ranging from -33 to -56 kJ/mol and -170 to -245 J/(mol K), respectively, when
referenced to a 1 Torr standard state for O2.30-45 The entropy change associated with dioxygen
binding is largely due to the loss of translational and rotational entropy by the bound oxygen as
predicted by statistical mechanical calculations.48
While it is expected that the affinity for dioxygen binding is strongly dependent on
porphyrin species, coordination of the cobalt ion by a fifth ligand can strongly influence
equilibrium. Stynes et al reported on the effects of coordination of basic ligands to the fifth
position on dioxygen binding to cobalt(II) protoporphyrin IX dimethyl ester below -30°C.30,31
Variation in enthalpy and entropy of oxygenation up to 24% and 15%, respectively, were
achieved by changing the fifth ligand. The same study also investigated Hammet relationships
among a series of para-substituted pyridines and concluded that oxygen binding is favored by
5
ligands capable of donating electron density to the metal ion to help offset the loss of electron
density associated with binding oxygen.30 Collman et al have also investigated the effect of
coordination of a fifth ligand on oxygenation of cobalt “picket fence” porphyrin at room
temperature.43,45 They showed that cobalt “picket fence” porphyrin binds oxygen to a degree
comparable to cobalt substituted myoglobins so long as an imidazole was also bound to the
cobalt (in the 5th coordination site). In the absence of a bound imidazole the porphyrin did not
bind oxygen at room temperature, thereby underscoring the importance of the fifth coordination
site. We also note that Summers and Stolzenberg indicated that ligand binding to the 5th
coordination site on cobalt(II) porphyrins relieved strain within the ring system.49
Reversible dioxygen binding to cobalt porphyrins and phthalocyanines on surfaces has been
reported by groups researching gas sensors23,24 and gas selective membranes.25,50,51 Cobalt(II)
“picket fence” porphyrin affixed to imidazole and pyridine bases imbedded in membranes are
capable of reversibly binding dioxygen with half oxygen saturation pressures (P1/2) lower than in
solution or as a crystal. The P1/2 of Co “picket fence” porphyrin in toluene with 1-
methylimidazol coordinated at the fifth site is 140 Torr at 25°C,45 while the same porphyrin
bound to an imidazol-terminated polymer reaches half saturation at 74 Torr at 25°C;51 close to
the P1/2 of crystalline porphin/1-methylimidazol (61 Torr) at 25°C.45 The enthalpy of
oxygenation for the porphyrin/imidazol pair increases in the order solution (-51.0 kJ/mol)45 to
solid state (-55.6 kJ/mol)45 to the membrane bound porphyrin (-58.6 kJ/mol).51 The entropy
change, referenced to a 1 Torr standard state, upon oxygen binding is less dramatic in this case.
The values increase from -214 J/(mol K-mol in toluene solution43 to -223 J/(mol K) in the solid
state. 45 Summers et al state that the entropy change for the membrane51 is similar to that in
solution.
6
The subject of this study, CoOEP, is a known electrocatalyst for the reduction of
oxygen27,52 and shows promise as a gas sensor due to its capability to bind volatile organic
molecules.53,54 Interestingly, Yamazaki et al have demonstrated that substrate choice can impact
the onset potential of oxygen reduction by cobalt porphyrins by almost 200 mV.27 CoOEP
reversibly binds carbon monoxide at ambient temperatures,55 but will bind dioxygen only at low
temperatures (<-90°C).56-59 The dioxygen complex was investigated by Raman, infrared,
Mössbauer, and electron spin resonance spectroscopies. In all but one of the studies an axial
base was used. In that latter study, the dioxygen adduct was reported to form in argon matrices
at 15 K without the addition of a fifth ligand.56 No thermodynamic data for adduct formation
was reported.
EXPERIMENTAL
2,3,7,8,12,13,17,18-Octaethyl-21H,23H-porphine cobalt(II) [CoOEP] was purchased
from Aldrich. Reagent grade chloroform was purchased from J.T. Baker. Phenyloctane (99%)
was purchased from Alfa Aesar. All reagents were used without further purification. 1 cm2
highly ordered pyrolytic graphite (HOPG) substrates (grade 2) were purchased from SPI supplies
(West Chester, Pa. #436HP-AB, lot #1160321). STM images were recorded using a Molecular
Imaging (now Agilent) Pico 5 STM equipped with a 1 µm2 head and an environmental chamber.
STM tips were made by cutting or electrochemically etching Pt0.8Ir0.2 wire (California Fine Wire
Company Grover Beach, Ca.). Images were typically obtained at a sample potential of -0.5 volts
and a tunneling current of 20 pA. Scan rates typically were 4.7 lines/sec, giving a total image
time of 2.0 minutes. The temperature of the sample was controlled by either a variable
temperature hot stage or a 1X Peltier stage using a Lakeshore 330 auto-tuning temperature
controller. Temperature was monitored with a calibrated Pt resistance thermometer. Stock
7
solutions of CoOEP were prepared by dissolving sufficient solid CoOEP in chloroform to make a
solution of concentration 200 M. This solution was subsequently diluted to about 9 µM in
chloroform. STM samples were prepared by dropping a 25 µL aliquot of the 9 µM CoOEP
solution on a freshly peeled 1 cm2 HOPG substrate and allowing the solvent to evaporate. The
dried surface was then covered with phenyloctane and the resulting sample was then placed on
the heating stage and transferred into the controlled atmosphere chamber of the STM. It should
be noted that samples prepared from pure chloroform in this way showed not structured
adsorbates and appeared to be clean graphite. This is worth noting since the weight/weight ppm
of CoOEP in chloroform is 3.6 ppm and the solid residue specification for the reagent used is < 1
ppm.
The partial pressure of oxygen over the sample was controlled by flowing mixtures of
argon and oxygen gas into the STM chamber. Oxygen and argon flows were controlled by
individual flow meters (model 7200 King Instrument Company Garden Grove, Ca.). Partial
pressures were determined as the ratio of O2/(O2+Ar) times the measured barometric pressure.
The gasses used had the following purity: 99.995% Argon with water <5 ppm, O2 <5 ppm, and
THC <2 ppm; 99.995% Oxygen with water <5 ppm, and THC <2 ppm; and, N2 with water < 5
ppm, O2 < 5 ppm, and THC < 2 ppm.
The sample (in the STM chamber) was first annealed at 100°C for 10 min under pure
argon flowing at 2.5 standard cubic feet per hour (SCFH). Prior to recording images the samples
were allowed to equilibrate for at least three hours at the oxygen partial pressure and temperature
of interest. All images were recorded while scanning in a drop of phenyloctane. All STM
images were background subtracted using SPIP60 image processing software and drift corrected
using a linear drift correction algorithm.61,62
8
Solution phase oxygen binding studies were carried out using a Perkin Elmer model 330
spectrophotometer with 1 cm pathlength cuvettes. All samples for solution phase oxygen
binding experiments were prepared in reagent grade toluene degassed by boiling for 45 minutes
followed by cooling to room temperature. The toluene was continuously purged with nitrogen
during the degassing process. A 1 µM solution of CoOEP prepared using the degassed toluene
was split into two equal samples. Oxygen gas was bubbled through one of the samples for 24
hours. After oxygen exposure the CoOEP solution was analyzed by UV-visible spectroscopy. A
duplicate control sample was exposed to nitrogen for 24 hrs and subsequently analyzed by UV-
visible spectroscopy. Appropriate blanks were prepared by exposing toluene to either oxygen or
nitrogen gas.
RESULTS AND DISCUSSION
Figure 1 is the structure of CoOEP in the all-up or crown configuration of the ethyl groups.
This is the known orientation of the ethyl groups for NiOEP adsorbed on Au(111)63 and is the
presumed configuration for CoOEP on HOPG. The space filling models used later are based on
optimization of this configuration by DFT calculations using the B3LYP functional and the 6-
311++G(d,p) basis. The actual optimized structure is provided as a Gaussian03 com file in the
supplemental materials. We also used the B3LYP functional and a much larger basis (6-
311++G(d,p) on C, N, and H and 6-311+G(3df) on O and Co, to estimate the structure of the O2-
CoOEP adduct.
To confirm previous results that pure CoOEP will not bind O2 in solution at room
temperature, we measured UV-Visible spectra of O2 and N2 saturated solutions in toluene after
9
24 hours of exposure. Figure 2 demonstrates that there is no change in either the Soret or Q
bands, indicating that no adduct forms under these conditions.
Figure 1: Model of “crown” configuration of CoOEP derived from DFT calculation (top and
side views). Black atoms are carbon, white are hydrogen, blue are nitrogen, and red are cobalt
The Gaussian03 com file for this structure is available as supplemental material.
10
Figure 3 presents a representative image obtained from the HOPG supported CoOEP surface
covered in phenyloctane and at equilibrium with a pure O2 atmosphere at 25ºC. There are
clearly three types of sites in the image. Those enclosed by squares are very deep and are
vacancies in the monolayer. Those enclosed by circles have apparent heights about half that of
the other molecular features. By performing the same measurements on monolayers annealed at
100ºC in pure argon, it is clear that the bright (high) features are simply CoOEP. This is
consistent with a number of previous studies wherein it was demonstrated that tunneling through
the half filled dz2 orbital produces the bright molecular center.64,65,66
Figure 2. UV-Vis spectrum of
CoOEP in toluene solution after
24 hours exposure to O2 and to
N2, respectively. The
oxygenated solution data is
manually offset upward to aid
comparison.
11
Figure 3: Drift corrected constant current STM image of the phenyloctane - CoOEP/HOPG
interface under conditions of O2 saturation at 25C. STM data was acquired at -0.5 V and 20 pA
setpoint. Note that the molecules enclosed in circles are considerably dimmer than others, and
the squares identify vacancies in the monolayer. The inset is a cross section emphasizing the
differences between the bright and dim molecules.
12
The components of a mixed monolayer of cobalt (II) hexadecafluoro-29H,23H-
phthalocyanine and nickel (II) tetraphenyl 21H,23H-porphine can easily be differentiated in
STM images by the bright half-filled dz2 orbital in the cobalt species as opposed to the darker Ni
atoms.65 The STM is also capable of detecting axial groups attached to porphyrins and
phthalocyanines. Vanadyl phthalocyanine centers appear dark in STM images due to oxygen’s
lack of states near the Fermi level.66,67 The striking difference in the STM contrast of bright
cobalt with dark vanadyl octaethyl porphyrins adsorbed on graphite was also well demonstrated
by Miyake et al.68 It is the STM’s ability to differentiate the bright half filled dz2 orbital in cobalt
coupled with the expected attenuated signal in the oxygenated species which allow oxygenated
and deoxygenated cobalt porphyrins to be distinguished and consequently thermodynamic data
to be gathered on a molecule by molecule basis.
In order to verify our interpretation of the bright and dark features, STM images at varying
equilibrium partial pressures of O2 were measured and analyzed. Defining the surface coverage
of dark molecules, , as the number of dark molecules divided by the total number of surface
molecules, one would expect the quantity /(1-) to be proportional to the partial pressure of O2
provided that the dark molecules were sites of O2 adsorption and the Langmuir isotherm is
followed. Figure 4 clearly shows that this is the case. Thus, the equilibrium constant for the O2
adsorption process, KP, is given by
0)1(
P
PKP , where following the convention in
solution phase studies we have taken the standard state to be P0 = 1 Torr. Alternatively, one
might relate the equilibrium constant to the molal concentration of O2 in solution, Kc, where
13
0)1(
c
cKc , and c0 is the hypothetical ideal state of 1 molal of O2 in solution. Using
GP,c = -RTln(KP,c), one can determine the free energy change for the
O2 + CoOEP/HOPG = O2-CoOEP/HOPG
system in the appropriate standard state. cP
cP T
GS
,,
for the appropriate standard state,
and HP,c = GP,c + TSP,c , thereby allowing all the thermodynamic functions to be obtained
from a series of STM measurements at various temperatures. The two sets of thermodynamic
quantities (at fixed c or fixed P) are connected by the temperature dependent Henry's law
constant. The thermodynamics that are derived from their temperature dependences differ by the
heat and entropy of solution of O2 in phenyloctane.
Before proceeding with the thermodynamic analysis, it is important to ascertain the
reliability in the values of obtained. Thus, the value used for each temperature was the
average of a large number of separate measurements of STM images. The standard deviations in
these measurements were used to determine the vertical error bars (± 1 standard deviation) in
each figure. To ensure that the system had come to equilibrium, these were measured after a
considerable wait and over a long period of time. A representative set of such measurements is
displayed in Figure 5. The data in Figure 5 was collected after the system had been allowed to
come to equilibrium for 4 hours. We note that each image analyzed contained about N = 250
molecules. However, at 25C and 176 Torr, the average coverage of oxygenated species is only
about 7% of the 250 molecules or 18 oxygenated molecules, resulting in an expected statistical
fluctuation of (1/N)1/2, or 23%, which is of the order observed in Figure 5.
14
Figure 4. Langmuir
plot of relative surface
coverage of dark
molecules as a function
of O2 partial pressure
at 25ºC.
Figure 5:
Variation in
measurements
of with time
for P(O2)= 176
Torr and 25ºC.
In order to determine K(T), values of were extracted from STM images acquired at
various temperatures and 100% O2 saturation and 704 Torr. Figure six shows three of the many
15
images analyzed as a function of temperature between 10ºC and 40ºC. Data were not acquired
above 40ºC because the number of oxygenated sites had become too few to have statistical
significance. It should be noted that the appearance of the bright center varied with tip quality.
For a very sharp tip the bright region was small and intense. For a dull tip, most of the molecule
appeared bright. In all cases it was easy to distinguish the dim and bright molecules within a
given image. It should be further noted that the “dim” molecules can also be distinguished from
vacancies in the monolayer. The central image in Figure 6 shows a cross section through three
bright molecules, one dim one, and a true vacancy (surrounded by a square).
Figure 6: Representative constant current STM data acquired at 10ºC (left), 25ºC (middle), and 40ºC (right).
Using average values and standard deviations for determined in the manner above for
100% O2 atmosphere (P = 704 Torr) at various temperatures, one can determine values of GP as
16
a function of temperature, as shown in Figure 7. From the slope of that data with temperature,
S may be determined. With these, one may calculate H. Thus, one arrives at HP = -68±10
kJ/mol and SP= -297±30 J/(mol K).
One may also report thermodynamic values referenced to the hypothetical ideal 1 molal (of
O2) state. These are denoted by subscript c. Thus, HcP GGG ,
HcP HHH , and HcP SSS , where the H subscript refers to the
change in thermodynamic quantity with respect to formation of the solution from the ideal gas
phase (Henry’s law). While the Henry's law constant for phenyloctane is not available as a
function of temperature, values for two closely related compounds, toluene and octane, are
available.69,70 Thus, one can estimate Hc and Sc based on the range seen for these solvents.
Hc(toluene) = -72±10 kJ/mol, and Sc(toluene) = -211±30 J/(mole K), while Hc(octane) =
-68±10 kJ/mol, and Sc(octane)= -204±30 J/(mole K). Because H for solvation of O2 is small
in these solvents, there is no significant effect on Hc and we can reasonably say that Hc
(phenyloctane) = -70±15 kJ/mol. While S for solvation of O2 is large, it is essentially the same
for both these solvents. Thus, it is reasonable to estimate Sc (phenyloctane) = -208±24 J/(mole
K). Because it is common to do so in the solution phase oxygen binding literature, we used our
data to determine the P1/2(298K), the O2 partial pressure at which half the sites on the
CoOEP/HOPG surface would be occupied by O2. We found that P1/2(298 K) = 3,200 Torr, or
about four atmospheres of oxygen.
These results are extremely surprising. Qualitatively, the reversible binding of O2 by
CoOEP/HOPG is a surprise. Based on solution phase studies, one does not expect to see
17
reversible binding of O2 by CoOEP at room temperature. Even the addition of pyridine bases to
the fifth coordination site is not sufficient to produce O2 binding at 298K. Nevertheless, the
HOPG bound CoOEP is binding oxygen and doing so with sufficient residence time to allow
observation by STM imaging. Extending arguments derived from on solution phase studies such
as those of Collman 48,45 and Stynes,30,31 it appears that the HOPG surface is strongly donating
electrons to the cobalt center thereby stabilizing the polarized Co-O2 complex. Quantitatively,
the results provide a mixed message. The value of HP (-68 kJ/mole) is more negative than the
larger values reported for cobalt complexes in solution (about -60 kJ/mol), but is consistent with
the unusually large binding and long residence time observed for O2 on CoOEP/HOPG. The
Value of Sp [-297 J/(mol K)] is more negative than previous reports for solution and solid state
binding by cobalt complexes [about -230 J/(mol K)], and also slightly larger than the negative of
the absolute entropy for the O2 molecule at 1 Torr and 298 K calculated from statistical
mechanics (-268 J/mol K).
18
Figure 7: GP as a function of temperature. The slope of the solid blue line is the measured SP,
the entropy change relative to the 1 Torr standard state. For reference, a dotted line is shown
with slope equal to the SP based solely upon the entropy of O2 at 298K.
Since SP is based on the slope, it is the least accurate of the values determined and the
possibility of numerical error must be addressed. To that end, a line of slope -268 J/(mol K) and
Hp=-60 kJ/mol is drawn over the actual free energy data in Figure 7. In order for the correct
value of the slope to be -268 J/(mol K), it would require only a slight deviation from the least
squares best fit line and would fall within one standard deviation of all but one of the measured
values. We consider this very good agreement with the expectations derived from statistical
19
mechanics. We also note that the associated Hp is close to values determined for other room
temperature oxygen binders. One might argue that not all the rotational entropy of O2 is lost
because there is probably some hindered rotational motion over the 4-fold porphyrin pocket.
Where then might this additional entropy be lost? A possibility is that the O2 binding is inducing
some additional order in the solvent, the porphyrin ring, or both.
A possibility worth considering is that there is more than one type of adsorption/desorption
process and that the second (or more) process is occurring too fast to be followed by STM.
While we cannot eliminate this from consideration, the implication is that there is significantly
more bound O2 than we have observed. Since the amount seen with the STM is already higher
than expected, the existence of an unrelated fast mechanism appears less likely.
The large binding of O2 on CoOEP adsorbed on HOPG is not so surprising when viewed in
the context of recent UHV studies of metal porphyrins on surfaces. It was shown in the 1980s
that cobalt(II) tetraphenylporphyrin (CoTPP) supported on TiO2 powder catalyzes the reduction
of NO and NO2 in the presence of CO and H2 with high efficiency despite the fact that neither
the free standing CoTPP nor the TiO2 powder is reactive on its own.71 This was explained by
saying that electronic structure of the Co ion was modified by partial electron transfer from the
TiO2 support. Recently there have been a number of UPS, XPS, and STS studies suggesting that
metalloporphyrins couple electronically to metal surfaces. 72-74 A very recent systematic study of
metal tetraphenylporphyrins on silver clearly shows the surface behaving as a fifth coordination
site in a “surface trans effect”. 75 Thus, there is precedent in the literature for a substrate to act as
an electron donating fifth ligand on the metal ion of an adsorbed metal porphyrin.
20
A different facet of this study of interest to surface science is the structure of the monolayer
of CoOEP. Figure 8 shows a portion of a drift corrected STM image and one possible set of
molecular orientations that generate the observed surface pattern. Unlike the UHV study of
NiOEP, we could not resolve the ethyl groups.63 Thus, we were forced to rely on simple model
building to arrive at a possible structure. Correctly scaled models (based on the DFT calculated
geometry), were placed over the image in a manner that appeared to minimize overlap of ethyl
groups. As shown in Figure 8, this resulted in a body centered unit cell at 25C having A =
2.92±0.15 nm, B = 2.73±0.15 nm, and = 54±3º. This cell results from a relative rotation of 18º
between molecules along the A direction. If one ignores the attempt to optimize packing and
simply uses the molecular pattern, a smaller lattice having only one molecule per unit cell results
with A' = 1.41±0.1 nm, B' = 1.36±0.1 nm, and = 54±4º. These value can be compared to those
reported for CoOEP/HOPG deposited from 1-tetradecene by Miyake et al. 68 They gave A'=
1.48±0.1, and B'=1.42±0.1, in reasonable agreement with the low resolution pattern. In the
absence of high resolution images, any integer multiple (nA', mB') of A' and B' might be the
actual surface structure.
21
Figure 8: Possible unit cell parameters for CoOEP/HOPG in phenyloctane at 25C. A =
(1.46±0.1 nm)n, B = (1.36±0.1 nm)m, and =54.3±4º.
CONCLUSIONS
It has been demonstrated that variable temperature STM can be used to determine adsorption
isotherms and thermodynamic data for processes occurring at the solid-solution interface. We
have shown that the solid support (in this case HOPG) can act in a manner similar to an electron
donating ligand bound to the fifth coordination site on the cobalt ion of CoOEP, thereby greatly
22
increasing the compound’s affinity for oxygen. The free energy, enthalpy, and entropy change
associated with the O2 binding process are determined for the first time and found to be
qualitatively correct but larger than previously observed for purely solution phase reactions.
While this difference may be due to surface and solvent reorganization, certain determination
will require extending the temperature range of these measurements to lower temperatures,
which we plan to do. Another obvious extension of this work is the measurement of the binding
process on substrates having different chemical properties and work functions. This is also
underway.
This work presents a novel approach for obtaining thermodynamic quantities and a better
understanding of the fundamental chemistry of reactions of dioxygen with metal organic
complexes at the single-molecule level in solution environments. The experimental conditions
used here allow one to mimic the chemical events of biological oxidation and in-situ
heterogeneous catalysis. Using a metal substrate as the 5th ligand allows for the formation of
transition metal-dioxygen binding complexes, where the bound dioxygen is held in a sterically
protected site, which in principle precludes dimerization and other unwanted reactions that may
occur in a solution phase.
ASSOCIATED CONTENT
Supporting Information. Gaussian com file for the optimized geometry shown in Figure 1.
This material is available free of charge via the Internet at http://pubs.acs.org.
ACKNOWLEDGEMENTS
This material is based upon work supported by the National Science Foundation under grants
CHE-1152951, CHE-1112156, and CHE-1058435.
23
1 Askadskaya, I.; Rabe, J. P. Phys. Rev. Let. 1992, 69, 1395-1398.
2 Walch, H.; Maier, A.; Heckl, W.; Lackinger, M. J. Phys. Chem. C 2009, 113, 1014–1019.
3 Ramachandran, G. K.; Hopson, T. J.; Rawlett, A. M.; Nagahara, L. A.; Primak, A.; Lindsay, S.
M. Science 2003, 300, 1413-1416.
4 Marie, C.; .Silly, F.; Tortech, L.; Mullen, K.; Fichou, D. ACS Nano 2010, 4, 1288-1292.
5 Giesen, M.; Baier, S. J. Phys.: Con. Mat. 2001, 13, 5009-5026.
6 MacLeod, J.; Rosei, F. Aust. J. Chem. 2011, 64, 1297-1298.
7 Dretschkow T.; Wandlowski, T.; Wandelt, K.; Thurgate, S. (Eds.): Solid–Liquid Interfaces,
Topics Appl. Phys. 2003, 85, 259–324. Springer-Verlag Berlin Heidelberg.
8 English, W.; Hipps, K. W. J. Phys. Chem. C 2008, 112, 2026-2031
9 Gutzler, R.; Sirtl, T.; Dienstmaier, J. F.; Mahata, K.; Heckl, W. M.; Schmittel, M.; Lackinger,
M. J. Amer. Chem. Soc. 2010, 132, 5084-5090.
10 Lindsay, S. M. (Molecular Imaging Corporation, USA). U.S. 1997, 11 pp. CODEN:
USXXAM US 5654546 A 19970805 Patent written in English. Application: US 95-551836
19951107.
11 Oulevey, F.; Gremaud, G.; Kulik, A. J.; Guisolan, B. Rev. Sci. Instrum. 1999, 70(3), 1889-
1890.
12 Trawick, M. L.; Angelescu, D. E.; Chaikin, P. M.; Valenti, M. J.; Register, R. A. Rev. Sci.
Instrum. 2003, 74(3, Pt. 1), 1390-1392.
13 Sattin, B.. D.; Goh, M. C. Rev. Sci. Instrum. 2004, 75(11), 4778-4780.
14 Campbell, W.M.; Jolley, K.W.; Wagner, P.; Wagner, K.; Walsh, P.J.; Gordon, K.C.; Schmidt-
Mende, L.; Nazeerudding, M.K.; Wang, Q.; Ratzel, M.; Officer, D.L. J. Phys. Chem. C Lett.
2007, 111, 11760-11762.
24
15 Shi, Z.; Zhang, J.; J. Phys. Chem. C 2007, 111, 7084-7090.
16 Yao, C.; Yan, L.; Guan, L.; Liu, C.; Song, P.; Su, Z. Dalton Trans, 2010, 39, 7645-7649.
17 Thyagariajan, S.; Leiding, T.; Arskold, S.P.; Cheprakov, A.V.; Vinogradov, S.A. Inorg. Chem.
2010, 49, 9909-9920.
18 Riberio, S.M.; Serra, A.C.; Rocha Gonsalves, A.M.d’A. J. Mol. Catal. A, 2010, 326, 121-127.
19 Verma, A.; Facchina, S.L.; Hirsch, D.J.; Song, S.; Dillahey, L.F.; Williams, J.R.; Snyder, S.H.
Molecular Medicine 1998, 4, 40-45.
20 Hu, J.S.; Ji, H.X.; Wan, L. J. Phys. Chem. C 2009, 113, 16259-16265.
21 Friesen, B.A.; Wiggins, B.; McHale, J.L.; Mazur, U.; Hipps, K.W. J. Amer. Chem. Soc. 2010,
132, 8554-8556.
22 Momenteau, M.; Reed, C.A.; Chem. Rev. 1994, 94, 659-698.
23 Bohrer, F.I.; Sharoni, A.; Colesniuc, C.; Park, J.; Schuller, I.K.; Cummel, A.C.; Trogler, W.C.
J. Amer. Chem. Soc. 2007, 129, 5640-5646.
24 Bohrer, F.I.; Colesniuc, C.N.; Park, J.; Ruidiaz, M.E.; Schuller, I.K.; Kummel, A.C.; Trogler,
W.C. J. Amer. Chem. Soc. 2009, 131, 478-485.
25 Hyakutake, T.; Okura, I.; Asai, K.; Nishide, H. J. Mater. Chem. 2008, 18, 917-922.
26 Su, B.; Hatay, I.; Trojanek, A.; Samec, Z.; Khoury, T.; Gros, C.P.; Barbe, J.M.; Daina, A.;
Carrupt, P.A.; Girault, H.H. J. Amer. Chem. Soc. 2010, 132, 2655-2662.
27 Yamazaki, S.; Yamada, Y.; Ioroi, T.; Fujiwara, N.; Siroma, Z.; Yasuda, K.; Miyazaki, Y. J.
Electroanal. Chem. 2005, 576, 253-259.
25
28 Steiger, B.; Anson, F.C. Inorg. Chem. 2000, 39, 4579-4585.
29 Yu, H. .Z.; Baskin, J.S.; Steiger, B.; Anson F.C.; Zewali, A.H. J. Amer. Chem. Soc. 1999, 121,
484-485.
30 Stynes, D.V,; Stynes, H.C.; James, B.R.; Ibers, J.A. J. Amer. Chem. Soc. 1973, 95, 1796-1801.
31 Stynes, H.C.; Ibers, J.A.; J. Amer. Chem. Soc. 1972, 94, 5125-5127
32 Beugelsdijk, T.J.; Drago, R.S. J. Amer. Chem. Soc. 1975, 97, 6466-6470.
33 Linard, J.E.; Ellis Jr., P.E.; Budge, J.R.; Jones, R.D.; Basolo, F. J. Amer. Chem. Soc. 1980,
102, 1896-1904.
34 Walker, F.A. J. Amer. Chem. Soc. 1973, 95, 1154-1159.
35 Yamamoto, H.; Takayanagi, T.; Kwan, T.; Yonetani, T. Bioninorg. Chem. 1977, 7, 189-201.
36 Stynes, H.C.; Ibers, J.A. J. Amer. Chem. Soc. 1972, 94, 1559-15562.
37 Uruma, K.; Tsuge, K.; Sasaki, Y.; Imamura, T. Chem. Lett. 2005, 34, 474-475.
38 Nakayama, S,; Tani, F.; Matsu-ura, M.; Naruta, Y. Chem. Lett. 2002, 496-497.
39 Walker, F.A. J. Amer. Chem. Soc.. 1970, 92, 4235-4244.
40 Steiger, B.; Baskin, J.S.; Anson, F.C.; Zewail, A.H. Angew. Chem. Int. Ed. 2000, 39, 257-260.
41 Wang, M.Y.R.; Hoffman, B.M.; Shire, S.J.; Gurd, F.R.N. J. Amer. Chem. Soc. 1979, 101,
7394-7397.
42 Spilburg, C.A.; Hoffman, B.M.; Petering, D.H. J. Bio. Chem. 1972, 247, 4219-4223.
43 Collman, J.P.; Brauman, J.I.; Doxsee, K.M.; Halbert, T.R.; Suslick, K.S. Proc. Natl. Acad. Sci.
1978, 75, 564-568.
26
44 Zou, S.; Baskin, J.S.; Zewail, A.H. Proc. Natl. Acad. Sci. 2002, 99, 9625-9630.
45 Collman, J.P.; Brauman, J.I.; Doxsee, K.M.; Halbert, T.R.; Hayes, S.E.; Suslick, K.S. J. Amer.
Chem. Soc. 1978, 100, 2761-2766.
46 Li, J.; Oliver, A.G.; Ferraudi, G.; Lappin, A.G.; Scheidt, W.R. Inorg. Chem. 2010, 49, 2398-
2406.
47 Valiotti, A.B.; Abakumova, R.A. Rus. J. Appl. Chem. 2007, 77, 1623-1626.
48 Collman, J.P.; Brauman, J.I.; Suslick, K.S.; J. Amer. Chem. Soc. 1975, 97, 7185-7186.
49 Summers, J. S.; Stolzenberg, A. M. J. Amer. Chem. Soc. 1993, 15, 10559-10567.
50 Nishide, H.; Suzuki, T.; Nakagawa, R.; Tsuchida, E. J. Amer. Chem. Soc. 1994, 116, 4503-
4504.
51 Suzuki, Y.; Nishide, H.; Tsuchida, E. Macromolecules 2000, 33, 2530-2534.
52 Song, E.; Shi, C.; Anson, F.C. Langmuir 1998, 14, 4315-4321.
53 Capan, I. Tarimci, C.; Capan, R. Sensors and Actuators B 2010, 144, 126-130.
54 Akrajas, M.; Salleh, M.; Yahaya, M. Sensors and Actuators B 2002, 85, 191-196.
55 Schmidt, E.; Zhang, H.; Chang, C.K.; Babcock, G.T.; Oertling. W.A. J. Amer. Chem. Soc.
1996, 118, 2954-2961.
56 Urban, M.W.; Nakamoto, K. Inorg. Chim. Acta 1982, 61, 77-81.
57 Baumgarten, M.; Winscom, C.J.; Lubitz, W. Appl. Magn. Reson. 2001, 20, 1-36.
58 Sakai, H.; Maeda, Y.; Ogoshi, H.; Sugimoto, H.; Yoshida, Z. Chem. Lett. 1978, 353-356.
59 Bajdor, K.; Kincaid, J.R.; Nakamoto, K. J. Amer. Chem. Soc. 1984, 106, 7741-7747.
27
60 SPIP is a product of Image Metrology, Lyngsø Alle 3A, DK-2970 Hørsholm, Denmark.
61 Tong, W.; Xue, Y.; Zimmt, M.B.; J. Phys. Chem. C 2010, 114, 20783-20792.
62 Tong, W. Controlling Self-Assembled Monolayer Morphology with Dipolar Interactions:
Preparation, STM Image Analysis and Simulations of 1,5-(Disubstituted)-Anthracene
Monolayers. Ph.D. Thesis, Brown University, Providence, RI, 2010.
63 Scudiero, L.; Barlow, D. E.; Hipps, K. W. J. Phys. Chem. B 2002, 106, 996-1003.
64 Scudiero, L.; Barlow, D. E.; Hipps, K. W. J. Phys. Chem. B 2000, 104, 11899-11905.
65 Hipps, K.W.; Scudiero, L.; Barlow, D.E.; Cooke, Jr., M.P. J. Amer. Chem. Soc. 2002, 124,
2126-2127.
66 Barlow, D.E.; Hipps, K.W. J. Phys. Chem. B 2000, 104, 5993-6000.
67 Mazur, U.; Hipps, K.W.; Reichers, S.L. J. Phys. Chem. C 2008, 112, 20347-20356.
68 Miyake, Y.; Tanaka, H.; Ogawa, T. Coll. Surf. A Physiochem. Eng. Aspects 2008, 313-314,
230-233.
69 Li, A.; Tang, S.; Tan, P.; Liu, C.; Liang, B. J. Chem. Eng. Data 2007, 52, 2339-2344.
70 Battino, R.; Rettich, T.R.; Tominaga, T. J. Phys. Chem. Ref. Data, 1983, 12, 163-178.
71 Mochida, I.; Tsuji, K.; Suetsugu, K.; Fujitsui, H.; Takeshita, K. J. Phys. Chem. 1980, 84,
3159-3162.
72 Bai, Y.; Sekita, M.; Schmid, M.; Bischof, T.; Steinruumlck, H. P.; Gottfried, J. M. Phys.
Chem. Chem. Phys. 2010, 12, 4336-4344.
73 Bai, Y.; Buchner, F.; Kellner, J.; Schmid, M.; Vollnhals, F.; Steinruumlck, H. P.; Marbach, H.;
Gottfried, J. M. New. J. Phys. 2009, 11, 125004-125019.
28
74 T. Lukasczyk, K. Flechtner, L.R. Merte, N. Jux, F. Maier, J.M. Gottfried, and H. P. Steinruck,
J. Phys. Chem. C. 2007, 111, 3090-3098.
75 Hieringer, W.; Flechtner, K.; Kretschmann, A.; Seufert, K.; Auwarter, W.; Barth, J. V.;
Gorling, A.; Steinruck, H. P.; Gottfried, J.M. J. Amer. Chem. Soc. 2011, 133, 6206-6222.
29
FIGURE CAPTIONS: Figure 1: Model of “crown” configuration of CoOEP derived from DFT calculation (top and
side views). Black atoms are carbon, white are hydrogen, blue are nitrogen, and red are cobalt
The Gaussian03 com file for this structure is available as supplemental material.
Figure 2. UV-Vis spectrum of CoOEP in toluene solution after 24 hours exposure to O2 and to
N2, respectively. The oxygenated solution data is manually offset upward to aid comparison.
Figure 3: Drift corrected constant current STM image of the phenyloctane - CoOEP/HOPG
interface under conditions of O2 saturation at 25C. STM data was acquired at -0.5 V and 20 pA
setpoint. Note that the molecules enclosed in circles are considerably dimmer than others, and
the squares identify vacancies in the monolayer. The inset is a cross section emphasizing the
differences between the bright and dim molecules.
Figure 4: Langmuir plot of relative surface coverage of dark molecules as a function of O2
partial pressure at 25ºC.
Figure 5: Variation in measurements of with time for P(O2)= 176 Torr and 25ºC.
Figure 6: Representative constant current STM data acquired at 10ºC (left), 25ºC (middle), and
40ºC (right).
Figure 7: GP as a function of temperature. The slope of the solid blue line is the measured
SP, the entropy change relative to the 1 Torr standard state. For reference, a dotted line is
shown with slope equal to the SP based solely upon the entropy of O2 at 298K
Figure 8: Possible unit cell parameters for CoOEP/HOPG in phenyloctane at 25C. A =
(1.46±0.1 nm)n, B = (1.36±0.1 nm)m, and = 54.3±4º.
30
FIGURE 1
31
FIGURE 2
32
FIGURE 3:
33
FIGURE 4
34
FIGURE 5
35
FIGURE 6
36
FIGURE 7
37
FIGURE 8
38
TOC Graphic